Communication systems employing digital transmitters and digital receivers are widely used. Such systems, which are commonly employed in mobile communication applications like cellular telephones, use digital modulation techniques such as binary phase shift keying (BPSK), quadrature phase shift keying (QPSK) or differential quadrature phase shift keying (DQPSK). Using these techniques, digital information is transmitted in bursts called frames, which are typically 20 milliseconds (ms) long. Frames generally have number of sections or subsections that may, for example, range in size from 160 to 640 microseconds (s). Each frame section typically contains numerous digital symbols that are transmitted approximately every 40 s. As is known, digital symbols may be encoded to each represent a number of digital bits. One frame section may be a preamble including a preamble bit sequence, which is known by each receiver that is to receive the transmitted frame, while other frame sections may include various bits representing digitized audio. In some applications, the preamble may be used to address the frame to a particular receiver or receivers. As will be appreciated by those familiar with the communication arts, not all of the information in each frame will be received error free because the fidelity of any particular communication system, while it may be high, is not perfect.
Bit error rate (BER) is a well-known metric that is used to specify or quantify the fidelity of a digital communication system. BER is a comparison between bits sent over a channel by a digital transmitter and bits received from the channel by a digital receiver. If the received bits are identical to the sent bits, the BER is zero, indicating that the communications system including the communication channel, the digital transmitter and the digital receiver has perfect fidelity. Conversely, if the bits received are substantially different from the bits that were sent, the communication system has low fidelity. For example, if there is one bit error in 100 bits, the BER is 0.01.
Digital communication systems are susceptible to various noise sources that decrease the fidelity of a communication system and, therefore, increase the BER of the communication system. Thermal noise (also called KT noise) is noise resulting from the temperature of various critical components in the digital communication system. Co-channel noise is noise caused by interference on the communication channel over which a digital transmitter is broadcasting. Of particular interest in mobile communication systems is multipath noise.
Multipath noise is noise caused by reception of delayed versions of a previously-received signal resulting from the fact that energy from a digital transmitter may take more than one path to a digital receiver. For example, energy from a digital transmitter that takes the most direct path to the receiver arrives at the receiver first, while energy taking another path, such as a path with one or more reflections from obstructions, the earth or the atmosphere, arrives at the digital receiver some relatively-short time later. Energy that does not take the most direct path from the digital transmitter to the digital receiver is called multipath energy, or simply “multipath.” In a mobile communications system, such as a cellular system, where one or both of a digital receiver and a digital transmitter are moving, the communication path between a transmitter and receiver is constantly changing and, therefore, so is the multipath. For example, as a person using a cellular phone travels in his or her car, multipath may range from nonexistent at one geographic location, to extremely high at another geographic location. Because the multipath is always changing, it is difficult for a digital receiver in a mobile system to combat the effects of multipath.
It is known to use an equalizing demodulator in a receiver to reduce the effects of multipath. An equalizing demodulator is a device that attempts to adapt a digital receiver to the characteristics of a channel to thereby minimize the effects of multipath before converting a received signal into a bit stream. Equalizing demodulators (commonly called equalizers) are actually non-equalizing demodulators that also perform computationally intensive equalization routines to equalize a channel. As a result, equalizing demodulators are typically slower and consume significantly more power than non-equalizing demodulators. Accordingly, it is known to use a non-equalizing demodulator to convert a received signal into digital bit stream when the fidelity of the communication system is high, and to use an equalizing demodulator when the fidelity of the communication system is low.
U.S. Pat. No. 5,283,531 to Serizawa et al. (hereinafter “Serizawa et al.”) discloses various techniques for selecting between an equalizing demodulator and a non-equalizing demodulator within a receiver. According to one technique, the Serizawa et al. system generates two bit streams, wherein one bit stream is generated by a non-equalizing demodulator and the other is generated by an equalizing demodulator. The fidelity of the bit streams from the non-equalizing demodulator and the equalizing demodulator are evaluated and the bit stream having the highest fidelity is selected for use in the receiver. Fidelity may be measured using bit stream coincidence with a reference bit stream or eye aperture measurements. A second technique disclosed in Serizawa et al. includes measuring fidelity using an eye aperture or a bit error rate of a bit stream generated by a non-equalizing demodulator and, based on a comparison to a threshold, selecting a bit stream from either a non-equalizing demodulator or an equalizing demodulator for use by the receiver. A third technique described in Serizawa et al. selects an output bit stream for use in the receiver based on the presence or absence of multipath. This technique uses a matched filter to generate a signal, which is then compared to a threshold; wherein the period of time the generated signal is above the threshold generally indicates the presence or absence of multipath. If multipath is present, the system selects the equalizing demodulator bit stream while, if multipath is not present, the system selects the non-equalizing demodulator bit stream for use by the receiver.
A shortcoming of the techniques disclosed in Serizawa et al. is that they all require the continuous operation of a non-equalizing demodulator, even if the equalizing demodulator output is selected for use by the receiver. Operating a non-equalizing demodulator, even when the equalizing demodulator output is selected for use, requires additional power consumption because redundant processing is performed. Additionally, the Serizawa et al. system makes a bit error rate measurement after the RF signals are converted into a digital bit stream by the non-equalizing demodulator, which requires additional processing time (i.e., the time it takes the non-equalizing demodulator to convert the RF signal into a bit stream) when the equalizing demodulator output is being selected for use.